CN114730386A - Teacher data generation device, teacher data generation method, teacher data generation program, and storage medium - Google Patents
Teacher data generation device, teacher data generation method, teacher data generation program, and storage medium Download PDFInfo
- Publication number
- CN114730386A CN114730386A CN201980102255.XA CN201980102255A CN114730386A CN 114730386 A CN114730386 A CN 114730386A CN 201980102255 A CN201980102255 A CN 201980102255A CN 114730386 A CN114730386 A CN 114730386A
- Authority
- CN
- China
- Prior art keywords
- data
- section
- teacher data
- teacher
- attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
An information processing device (1) as a teacher data generation device generates teacher data used for data analysis using machine learning. An information processing device (1) is provided with: a section clipping unit (18) that acquires time-series data and clips, from the time-series data, data in a section that matches a pre-specified condition and is included in the time axis of the time-series data, thereby generating data for each section; and a teacher data creation unit (19) that acquires attribute information that indicates the attribute of the data in each section and that is related to a pre-specified attribute, and creates teacher data by assigning a label that assigns the attribute information to the data in each section.
Description
Technical Field
The present invention relates to a teacher data generation device, a teacher data generation method, a teacher data generation program, and a storage medium that generate teacher data used for data analysis using machine learning.
Background
The teacher data is used for data analysis using teacher learning, which is one of machine learning. The teacher data is generated by applying a label to data to be analyzed, the label being information indicating data attributes. To enable highly accurate data analysis by teacher learning, generation of a large amount of teacher data is sought.
Patent document 1 discloses a device that extracts data that is the basis of teacher data from time-series data and applies a label to the extracted data. The device of patent document 1 presents data extracted from time-series data to a user. In the device of patent document 1, data to be tagged is selected by the user from the presented data. The device of patent document 1 performs labeling in accordance with an input operation performed by a user.
Patent document 1: japanese patent laid-open publication No. 2016-76073
Disclosure of Invention
In the conventional technique described in patent document 1, the user needs to determine the attribute of data for all presented data and manually apply a label to the data. Therefore, according to the related art, there is a problem that a large number of man-hours are required for creation of teacher data.
The present invention has been made in view of the above problems, and an object of the present invention is to provide a teacher data generating device capable of reducing the number of steps required for generating teacher data.
In order to solve the above problems, a teacher data generation device according to the present invention generates teacher data used for data analysis using machine learning. The present invention relates to a teacher data generation device, comprising: a section cutting unit that acquires time-series data and cuts data in a section that matches a predetermined condition included in a time axis of the time-series data from the time-series data to generate data for each section; and a teacher data creation unit that acquires attribute information that indicates an attribute of the data in each section and is related to a pre-specified attribute, and creates teacher data by assigning a label that assigns the attribute information to the data in each section.
ADVANTAGEOUS EFFECTS OF INVENTION
The teacher data generation device according to the present invention has an effect of reducing the number of steps required to generate teacher data.
Drawings
Fig. 1 is a diagram showing a configuration of a teacher data generation device according to embodiment 1 of the present invention.
Fig. 2 is a diagram for explaining status data acquired by the teacher data generating device according to embodiment 1.
Fig. 3 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 1.
Fig. 4 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 2 of the present invention.
Fig. 5 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 3 of the present invention.
Detailed Description
Next, a teacher data generation device, a teacher data generation method, a teacher data generation program, and a storage medium according to embodiments of the present invention will be described in detail with reference to the drawings. The present invention is not limited to this embodiment.
Embodiment mode 1
Fig. 1 is a diagram showing a configuration of a teacher data generation device according to embodiment 1 of the present invention. Fig. 1 shows an information processing apparatus 1 as a teacher data generation apparatus, and a peripheral device 2 and a control device 3 connected to the information processing apparatus 1. The information processing apparatus 1 generates teacher data. Teacher data is used for data analysis using machine learning. The information processing apparatus 1 generates teacher data using data input from the peripheral device 2 and the control device 3.
The control device 3 is a device that controls a control object. The control device 3 is a Controller for controlling a device such as a production apparatus or a plant apparatus, and is, for example, a Programmable Logic Controller (PLC). The control device 3 controls 1 or more devices as control objects. The control device 3 may be a controller other than a PLC or a numerical control device. The control device 3 may be a controller that controls the production apparatus or a device other than the apparatus. The teacher data generated by the information processing device 1 is used for analyzing data indicating the state of the control target.
The information processing apparatus 1 is an apparatus that processes data regarding a control target, and is, for example, a display. The display performs data processing for displaying information on the state of the control object, and data processing for generating teacher data. The information processing apparatus 1 may be an apparatus other than a display. The information processing apparatus 1 may be a computer such as a personal computer. The information processing device 1 is provided with a teacher data generation program that is a program for generating teacher data.
An arbitrary number of peripheral devices 2 are connected to the information processing apparatus 1. The peripheral device 2 transmits data indicating the state of the control target to the information processing apparatus 1. The peripheral device 2 includes, for example, a sensor for detecting an operation state of a control target. The peripheral device 2 may include a device provided outside the control target, or may include a device provided inside the control target. Further, data indicating the state of the control target may be transmitted from the control device 3 to the information processing apparatus 1.
The control device 3 stores control information, which is various information on the control of the control object. The control device 3 transmits control information designated in advance among the control information to the information processing apparatus 1. The information processing apparatus 1 uses the control information transmitted from the control device 3 as attribute information indicating an attribute of status data described later. The control device 3 transmits control information to the information processing apparatus 1 in accordance with a request issued by the information processing apparatus 1. The control device 3 may transmit the control information to the information processing apparatus 1 in accordance with a determination made by the control device 3, not in accordance with a request issued by the information processing apparatus 1.
The information processing apparatus 1 includes an input device 11 for inputting information, a communication device 12 for performing communication with the peripheral device 2 and communication with the control device 3, a display device 13 for displaying information, a processor 14 for executing various processes, and a memory 15 for storing information.
The input device 11 is a device such as a keyboard, a mouse, or a touch panel. The communication device 12 is a connection interface with an external device of the information processing device 1. The display device 13 displays information on a screen.
The processor 14 is a CPU (Central Processing Unit). The processor 14 may be a processing device, an arithmetic device, a microprocessor, a microcomputer, or a dsp (digital Signal processor). The Memory 15 includes RAM (random Access Memory), ROM (Read Only Memory), flash Memory, EPROM (Erasable Programmable Read Only Memory) or EEPROM (registered trademark), HDD (hard Disk drive) or SSD (solid State drive). The teacher data generation program is stored in the memory 15. The processor 14 executes programs stored in the memory 15.
The functional structure implemented by using the processor 14 is shown in fig. 1. The data processing unit 16 performs data processing for generating teacher data. The display processing unit 17 performs processing for displaying on the display device 13. The functions of the data processing unit 16 and the display processing unit 17 are realized by a combination of the processor 14 and software. The functions of the data processing unit 16 and the display processing unit 17 may be realized by a combination of the processor 14 and firmware, or may be realized by a combination of the processor 14, software, and firmware. The software or firmware is described as a program and is stored in the memory 15. The processor 14 reads the software or firmware. Processor 14 executes software or firmware.
The teacher data generation program may be recorded on a storage medium that can be read by a computer. The information processing device 1 may store the teacher data generation program recorded in the storage medium in the memory 15. The storage medium may be a floppy disk, i.e., a portable storage medium, or a semiconductor memory, i.e., a flash memory. The teacher data generation program may be installed in the information processing apparatus 1 from another computer or server apparatus via a communication network.
Information is input to the input device 11 by an operation performed by an operator. The communication device 12 receives data transmitted from the peripheral device 2 or the control device 3 and control information transmitted from the control device 3. The communication device 12 outputs the received data to the display processing unit 17. The memory 15 stores data received by the communication device 12. By accumulating data received by the communication device 12 in the memory 15 as needed, the state data, which is time-series data representing the state of the control target, is stored in the memory 15.
The display processing unit 17 generates display information such as character information or image information based on the content of the state data. The display processing unit 17 outputs the generated display information to the display device 13. The display device 13 displays the display information.
The data processing unit 16 includes a section clipping unit 18 and a teacher data creation unit 19. The section clipping unit 18 acquires the state data from the memory 15. The section clipping unit 18 clips data in a section corresponding to a predetermined condition included in the time axis of the time-series data from the time-series data, thereby generating data of each section. The memory 15 stores data of each section. In the following description, data of each section is sometimes referred to as section data. The operator operates the input device 11 to specify conditions for specifying the section.
The communication device 12 outputs the received attribute information, which is the control information, to the teacher data creation unit 19. The teacher data creation unit 19 acquires the attribute information from the communication device 12. The attribute information is information indicating an attribute of the section data, and is information on a previously specified attribute. The attribute information is information equivalent to metadata regarding section data. The operator specifies the attribute by operating the input device 11.
The teacher data creating unit 19 acquires section data stored in the memory 15. The teacher data creation unit 19 may directly acquire section data from the section clipping unit 18. The teacher data creation unit 19 creates teacher data by assigning a label of attribute information to the section data. The memory 15 stores the created teacher data.
In the following description, the control target is a processing machine that processes a workpiece. The teacher data generated by the information processing device 1 is used for data analysis for preventive maintenance of the processing machine. The peripheral device 2 detects the state of the processing machine and transmits data, which is the detection result, to the information processing apparatus 1. The peripheral device 2 uses various sensors for observing the state of the processing machine. Examples of the various sensors include a current sensor, a voltage sensor, a vibration sensor, and a temperature sensor. The peripheral device 2 may use a camera that captures the state of processing performed by the processing machine. The control information transmitted from the control device 3 to the information processing apparatus 1 includes information related to processing, such as information on processing conditions and information on processing time.
Fig. 2 is a diagram for explaining status data acquired by the teacher data generating device according to embodiment 1. Here, the state data is a current value of a current flowing through the servo motor. The servo motor drives a spindle of the processing machine. The memory 15 accumulates current values detected by the current sensors, which are the peripheral devices 2, to generate time-series status data. Fig. 2 shows a current waveform as a graph showing an example of the state data.
The state data is not limited to the current value as long as it is data indicating the state of the control target. The state data may be time-series image data obtained by a camera. The status data may be data describing the contents of processing executed by the control device 3.
The section cutting unit 18 cuts out state data in a section that meets a predetermined condition from the time-series state data. The section corresponding to the condition specified in advance is a section from a time point at which an event occurs in the action of the control target controlled by the control device 3 to an end of the event in the action of the control target. The section cutting unit 18 cuts state data in a period from a time point when an event occurs during the operation of the processing machine to an end of the event during the operation of the processing machine from the time-series state data.
Here, section data from the start of machining to the end of machining for each workpiece is specified in advance. The section cutting unit 18 cuts section data in a section from the start of machining of the workpiece to the end of machining of the workpiece from the state data for each workpiece. Thus, the section clipping unit 18 generates section data for each cycle. The cycle may be any of a cycle related to a single process performed by 1 processing machine and a cycle related to a plurality of processes performed by a plurality of processing machines. The cycle may be a cycle per production line having a plurality of processing machines, or a cycle per factory having a plurality of production lines.
The section clipping unit 18 may clip the section data according to the event information. The event information is information indicating that an event has occurred. An event is any event related to an action of a control object. An example of the event information is an alarm notifying an abnormality of the control object. The control device 3 stops the machining when an alarm occurs. The section cutting unit 18 cuts section data triggered by an alarm. The event information, i.e., the alarm, is transmitted from the control device 3 to the information processing apparatus 1.
The event information may be a result of quality determination regarding the manufactured product. In this case, the section cutting unit 18 cuts the section data when the result of non-defective product determination is input. The event information may also be a product number of the product. In this case, the section cutting unit 18 cuts the section data when the product number of the product is changed. Further, the result of non-defective determination may be transmitted from the device that performs non-defective determination to the information processing apparatus 1. In this case, the peripheral device 2 includes a device for performing non-defective product determination. As described above, the event information is not limited to information obtained by transmission from the control device 3, and may be information obtained by transmission from the peripheral device 2.
Next, the attribute information will be explained. The attribute information is information set in advance regarding an operation performed by the control target or information indicating a result of the operation performed by the control target. Here, "processing conditions" are specified in advance as one of the attributes set as the attribute information. When the "machining condition" is designated, the attribute information on the "machining condition" includes parameters on the designated item among various items set at the time of machining, such as a machining path and a machining speed. The machining path is a moving path of the tool relative to the workpiece. The control information transmitted by the control device 3 contains attribute information, i.e., parameters, about "processing conditions". The parameter as the attribute information is information set in advance with respect to the operation performed by the processing machine. The teacher data creation unit 19 assigns parameters as attribute information to the section data.
The attribute to be used as the attribute information may be "processing time", "material lot number", "product number", "result of non-defective product determination", "OEE (Overall Equipment efficiency)", or "peak frequency", in addition to "processing condition". The "machining time" is a time from the start of machining of a workpiece to the end of machining of the workpiece, and is the above-described cycle. "Material lot" is a lot number that is appended to the material used in the process. The "product number" is the product number of the manufactured product. The "result of non-defective product determination" is a result of non-defective product determination regarding the manufactured product. "OEE" is an index indicating the overall efficiency of the plant, and is calculated based on each index of the operating rate, performance, and quality. OEE may be any of an index obtained by evaluation for 1 or more processing machines, an index obtained by evaluation for a production line, and an index obtained by evaluation for a factory. The "peak frequency" is the frequency of the peak included in the current waveform. The attribute information, which is information on the "machining time", "material lot number", "product number", "result of non-defective product determination", "OEE", and "peak frequency", is information indicating the result of the operation performed by the control target.
Further, the result of non-defective determination may be transmitted from the device that performs non-defective determination to the information processing apparatus 1. The data of OEE may be transmitted from a device that calculates OEE to the information processing apparatus 1. As described above, the attribute information is not limited to the information transmitted from the control device 3, and may be information transmitted from a device other than the control device 3. A device that is the transmission source of the attribute information transmits control information to the information processing apparatus 1 in accordance with a request issued by the information processing apparatus 1. The device that is the transmission source of the attribute information may transmit the control information to the information processing apparatus 1 not in accordance with the request issued by the information processing apparatus 1 but in accordance with the determination made by the device that is the transmission source.
The attribute information given to the section data is not limited to the information about 1 attribute. The attribute information given to the section data may include information on a plurality of attributes. The operator can designate an arbitrary number of attributes as the attributes included in the attribute information. The operator can instruct the information processing apparatus 1 to replace the designated attribute by operating the input device 11. The teacher data creation unit 19 replaces the attribute included in the attribute information in accordance with the operation performed by the operator.
Fig. 3 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 1. The communication device 12 receives data transmitted from the peripheral device 2 or the control device 3. The memory 15 stores the status data by accumulating the data received by the communication device 12. In step S1, the section clipping unit 18 reads the state data from the memory 15 to acquire the state data. In step S2, the section clipping unit 18 clips the section data from the state data acquired in step S1. The memory 15 stores the section data generated in step S2.
In step S3, the teacher data creation unit 19 acquires the attribute information. The teacher data creating unit 19 acquires section data from the memory 15 or the section cutting unit 18. In step S4, the teacher data creation unit 19 creates teacher data by assigning a label of the attribute information to the section data. In step S5, the memory 15 stores the teacher data created in step S4. Thereby, the information processing apparatus 1 terminates the operation realized by the flow shown in fig. 3.
The section clipping unit 18 clips section data in a section corresponding to a predetermined condition from the state data according to the condition. The section clipping unit 18 can automatically create section data that is the basis of the teacher data by specifying conditions for clipping sections in advance.
The teacher data creation unit 19 acquires attribute information on the designated attribute from the control device 3 or the peripheral device 2. The teacher data creation unit 19 adds the acquired attribute information to the section data. The information processing apparatus 1 can automatically acquire attribute information and assign a label to the attribute information by the teacher data creation unit 19.
According to embodiment 1, the teacher data generation device includes the section clipping unit 18 and the teacher data creation unit 19, and can automatically generate section data that is a basis of the teacher data and assign a label to the section data. The generation of the section data and the labeling can be automatically performed without manual operation, and the teacher data generation device can reduce the man-hours required for generating the teacher data. Thus, the teacher data generation device has an effect of reducing the number of steps required to generate the teacher data.
Embodiment mode 2
Fig. 4 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 2 of the present invention. The configuration of the teacher data generation device according to embodiment 2 is the same as that of the teacher data generation device according to embodiment 1. In the information processing device 1 as the teacher data generation device according to embodiment 2, the teacher data creation unit 19 determines whether or not tagging is necessary based on information acquired for a preset attribute. In embodiment 2, the description will be mainly given of differences from embodiment 1.
The flow of steps S1 to S3 is the same as the case shown in fig. 3. In step S11, the teacher data creation unit 19 determines whether or not the section data generated by the cutting in step S2 is a target of the label assignment. The teacher data creation unit 19 acquires information for determining whether or not the label assignment is necessary, as in the case of the attribute information assigned to the section data. The teacher data creation unit 19 sets in advance an attribute for determining whether or not tagging is necessary, and a determination criterion as to whether or not tagging is necessary. The attribute for determining whether or not the tag is necessary to be added and the determination criterion for determining whether or not the tag is necessary are set by the operation of the input device 11 by the operator.
If the generated section data is not the target of the tagging (No at step S11), the information processing apparatus 1 terminates the operation realized by the flow shown in fig. 4. If the generated section data is the target of tagging (Yes at step S11), the information processing apparatus 1 advances the flow to step S4. The flow of steps S4 to S5 is the same as the case shown in fig. 3.
Here, specific examples of the case where the label is applied and the case where the label is not applied will be described. The attribute information given to the section data is set as a parameter of the machining speed. The memory 15 stores status data acquired when the product of product number "a" is manufactured and status data acquired when the product of product number "B" is manufactured. As an attribute for determining whether or not the label addition is necessary, the teacher data creation unit 19 sets a "product number". As a criterion for determining whether or not the labeling is necessary, it is set that the labeling is performed when the information acquired for the "product number" is information indicating the product number "a", and the labeling is not performed when the information acquired for the "product number" is information indicating the product number "B".
In the case of this specific example, the teacher data creation unit 19 assigns attribute information, which is a parameter of the processing speed, to the section data when the section data is the section data relating to the product number "a". The memory 15 stores teacher data, which is section data to which attribute information is added. When the section data is the section data relating to the product number "B", the teacher data creating unit 19 does not give the section data attribute information, which is a parameter of the processing speed. In this case, the teacher data is not created in the teacher data creating unit 19. Further, the operator can instruct the information processing apparatus 1 to change the attribute for determining whether or not the label is required to be applied and to change the determination reference by operating the input device 11.
According to embodiment 2, the teacher data creation unit 19 can dynamically change whether or not to apply a label based on information acquired for a preset attribute. The teacher data generation device can remove data unnecessary for data analysis at the time point when the teacher data is generated. Thus, the teacher data generation device can suppress the data amount of the teacher data to be generated.
Embodiment 3
Fig. 5 is a flowchart showing an operation flow of the teacher data generation device according to embodiment 3 of the present invention. The configuration of the teacher data generation device according to embodiment 3 is the same as that of the teacher data generation device according to embodiment 1. In the information processing device 1 as the teacher data generation device according to embodiment 3, the teacher data creation unit 19 compares the attribute information and adds information indicating the comparison result to the attribute information. In embodiment 3, the description will be made mainly on differences from embodiments 1 and 2.
In step S21, if the 1 st section data, which is the section data for 1 section, is input to the teacher data creation unit 19, the teacher data creation unit 19 determines whether or not there is teacher data, which includes the 2 nd section data, which is the section data acquired under the condition common to the 1 st section data, among the teacher data stored in the memory 15. The teacher data creation unit 19 determines that the 1 st section data and the 2 nd section data are section data acquired under a common condition when the attribute information added to the 1 st section data and the attribute information added to the 2 nd section data include information of the processing condition and the same information is included as information of the processing condition. The teacher data creation unit 19 may determine whether or not the 1 st section data and the 2 nd section data are section data acquired under a common condition based on information other than the information of the processing condition.
When determining that the teacher data of section 2 is not included in the teacher data stored in the memory 15 (No at step S21), the information processing device 1 ends the operation realized by the flow shown in fig. 5. If it is determined that the teacher data including the 2 nd section data is included in the teacher data stored in the memory 15 (Yes at step S21), the information processing apparatus 1 advances the flow to step S22.
In step S22, the teacher data creation unit 19 compares the information indicating the attribute of the 1 st section data with the information indicating the attribute of the 2 nd section data. As described above, when the teacher data including the 2 nd section data is created before the 1 st section data is created, the teacher data creating unit 19 compares the information indicating the attribute of the 1 st section data with the information indicating the attribute of the 2 nd section data.
Here, when the attribute information added to the 1 st section data and the attribute information added to the 2 nd section data include information of the "peak frequency", the teacher data creating unit 19 compares the value of the "peak frequency" of the 1 st section data with the value of the "peak frequency" of the 2 nd section data. The teacher data creation section 19 calculates a difference between the value of the "peak frequency" of the 1 st section data and the value of the "peak frequency" of the 2 nd section data. Since the "peak frequency" changes when the tool deteriorates, the difference in the "peak frequency" can be used as an index indicating the deterioration state of the tool.
In step S23, the teacher data creation unit 19 adds information indicating the comparison result in step S22 to the attribute information added to the 1 st section data. The teacher data creation unit 19 adds, as information indicating the comparison result, a value that is the calculation result of the difference value to the attribute information. Thereby, the information processing apparatus 1 terminates the operation realized by the flow shown in fig. 5.
The teacher data creating unit 19 sets in advance an attribute for comparing information. Further, the teacher data creation unit 19 may compare information on attributes other than the "peak frequency". The attribute for performing the comparison of the information is set by the operation of the input device 11 by the operator. The operator can instruct the information processing apparatus 1 to change the attribute for comparing the information by operating the input device 11.
In data analysis using teacher data, attribute information that is information indicating a comparison result can be used as an index indicating the credit of attribute information added to the teacher data. According to embodiment 3, the teacher data generation device can include information indicating the credit of the attribute information in the generated teacher data.
The configuration shown in the above embodiment is an example of the contents of the present invention, and may be combined with other known techniques, and a part of the configuration may be omitted or modified within a range not departing from the gist of the present invention.
Description of the reference numerals
The system comprises an information processing device 1, a peripheral device 2, a control device 3, an input device 11, a communication device 12, a display device 13, a processor 14, a memory 15, a data processing unit 16, a display processing unit 17, a section clipping unit 18 and a teacher data creation unit 19.
Claims (9)
1. A teacher data generation device for generating teacher data used for data analysis using machine learning,
the teacher data generation device is characterized by comprising:
a section clipping unit that acquires time-series data and clips, from the time-series data, data in a section that matches a predetermined condition included in a time axis of the time-series data, thereby generating data for each section; and
and a teacher data creation unit that acquires attribute information that indicates an attribute of the data in each section and is related to a pre-specified attribute, and creates the teacher data by assigning a label that assigns the attribute information to the data in each section.
2. The teacher data generating apparatus according to claim 1,
the teacher data creation unit determines whether or not the tagging is necessary based on information on a preset attribute.
3. The teacher data generating apparatus according to claim 1,
in the case where the teacher data including the 2 nd section data, which is the data of each section acquired under the condition common to the 1 st section data, is created before the 1 st section data, which is the data in 1 section, is created, the teacher data creating unit compares the information indicating the attribute of the 1 st section data with the information indicating the attribute of the 2 nd section data, and adds the information indicating the comparison result to the attribute information added to the 1 st section data.
4. The teacher data generating apparatus according to any one of claims 1 to 3,
the section cutting unit cuts the state data of each section from the state data that is the time-series data indicating the state of the control target controlled by the control device,
the teacher data creation unit creates the teacher data by assigning a label of the attribute information to the state data of each section.
5. The teacher data generating apparatus according to claim 4,
the section corresponding to the predetermined condition is a section from a time point at which an event occurs in the operation of the control target to an end of the event in the operation of the control target.
6. The teacher data generating apparatus according to claim 4 or 5,
the attribute information is information set in advance for an operation performed by the control target or information indicating a result of the operation performed by the control target.
7. A teacher data generation method generates teacher data used in machine learning by a teacher data generation device,
the teacher data generation method is characterized by comprising the following steps:
acquiring time sequence type data;
generating data of each section by cutting data in a section which is included in a time axis of the time-series data and which matches a predetermined condition from the time-series data;
acquiring attribute information indicating an attribute of the data of each section and relating to a previously specified attribute; and
the teacher data is created by assigning a label to the data of each section, the label being assigned with the attribute information.
8. A teacher data generation program for causing a computer to function as a teacher data generation device for generating teacher data used in machine learning,
the teacher data generation program causes the computer to execute the steps of:
acquiring time sequence type data;
generating data of each section by cutting data in a section which is included in a time axis of the time-series data and which conforms to a predetermined condition from the time-series data;
acquiring attribute information indicating an attribute of the data of each section and relating to a previously specified attribute; and
the teacher data is created by assigning a label to the data of each section, the label being assigned with the attribute information.
9. A storage medium storing the teacher data generation program according to claim 8 and readable by a computer.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2019/045302 WO2021100128A1 (en) | 2019-11-19 | 2019-11-19 | Training data generation apparatus, training data generation method, training data generation program, and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114730386A true CN114730386A (en) | 2022-07-08 |
Family
ID=73022493
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201980102255.XA Pending CN114730386A (en) | 2019-11-19 | 2019-11-19 | Teacher data generation device, teacher data generation method, teacher data generation program, and storage medium |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP6779411B1 (en) |
| CN (1) | CN114730386A (en) |
| WO (1) | WO2021100128A1 (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106339756A (en) * | 2016-08-25 | 2017-01-18 | 北京百度网讯科技有限公司 | Training data generation method and device and searching method and device |
| JP2017138509A (en) * | 2016-02-04 | 2017-08-10 | カシオ計算機株式会社 | Personal adaptation method for emotion estimator, emotion estimation device, and program |
| CN108279767A (en) * | 2017-01-05 | 2018-07-13 | 株式会社东芝 | Motion analysis device and action-analysing method |
| JP2018173865A (en) * | 2017-03-31 | 2018-11-08 | オムロン株式会社 | Control device |
| JP2019028836A (en) * | 2017-08-01 | 2019-02-21 | 三菱日立パワーシステムズ株式会社 | Learning assistance device for plant and learning assistance method for plant |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2019016035A (en) * | 2017-07-04 | 2019-01-31 | メトロエンジン株式会社 | Room charge setting device, room charge setting method, and program |
| JP6844562B2 (en) * | 2018-03-13 | 2021-03-17 | オムロン株式会社 | Annotation method, annotation device, annotation program and identification system |
| JP6892634B2 (en) * | 2018-03-15 | 2021-06-23 | オムロン株式会社 | Information processing equipment, information processing methods, and programs |
-
2019
- 2019-11-19 WO PCT/JP2019/045302 patent/WO2021100128A1/en active Application Filing
- 2019-11-19 JP JP2020520668A patent/JP6779411B1/en active Active
- 2019-11-19 CN CN201980102255.XA patent/CN114730386A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017138509A (en) * | 2016-02-04 | 2017-08-10 | カシオ計算機株式会社 | Personal adaptation method for emotion estimator, emotion estimation device, and program |
| CN106339756A (en) * | 2016-08-25 | 2017-01-18 | 北京百度网讯科技有限公司 | Training data generation method and device and searching method and device |
| CN108279767A (en) * | 2017-01-05 | 2018-07-13 | 株式会社东芝 | Motion analysis device and action-analysing method |
| JP2018173865A (en) * | 2017-03-31 | 2018-11-08 | オムロン株式会社 | Control device |
| JP2019028836A (en) * | 2017-08-01 | 2019-02-21 | 三菱日立パワーシステムズ株式会社 | Learning assistance device for plant and learning assistance method for plant |
Also Published As
| Publication number | Publication date |
|---|---|
| JP6779411B1 (en) | 2020-11-04 |
| WO2021100128A1 (en) | 2021-05-27 |
| JPWO2021100128A1 (en) | 2021-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110275506B (en) | Abnormality detection system, support device, and model generation method | |
| CN110275505B (en) | Abnormality detection system, support device, and model generation method | |
| US10452052B2 (en) | Method and system for processing machine data before completion of machining | |
| US20190188110A1 (en) | Industrial control system, and assistance apparatus, control assist method, and program thereof | |
| JP6603272B2 (en) | Information collecting apparatus and information collecting method | |
| JP2018173883A (en) | Control device, control program, control system, and control method | |
| JP7143639B2 (en) | Anomaly detection system, configuration tool device, and anomaly response function block | |
| JP2020052714A5 (en) | ||
| CN114730386A (en) | Teacher data generation device, teacher data generation method, teacher data generation program, and storage medium | |
| CN113874800A (en) | Support system for improving production efficiency | |
| TW201713456A (en) | Manufacturing efficiency optimization platform and tool condition monitoring and prediction method | |
| US11073820B2 (en) | Machining support device, numerical controller, and machining support system | |
| US20230129280A1 (en) | Computer readable storage medium, program creation support apparatus, and program creation support method | |
| CN113348415B (en) | Device state reproducing device, method and storage medium | |
| JP2020201982A (en) | Operation management device | |
| CN114981743B (en) | Design support device, design support method, and computer-readable storage medium | |
| WO2022162957A1 (en) | Information processing device, control system, and report output method | |
| US10935965B2 (en) | Operation management apparatus | |
| CN110879573A (en) | Machining time prediction device | |
| WO2023089773A1 (en) | Abnormality diagnostic device, abnormality diagnostic system, and storage medium | |
| EP4036752A1 (en) | Method for analysing process data related to a segment of a production process | |
| US11983003B2 (en) | Production efficiency improvement support system | |
| WO2023053511A1 (en) | Control system, information processing method, and information processing device | |
| JP2019003353A (en) | Method for analyzing operational state of control object device | |
| US20230367927A1 (en) | Non-transitory computer readable storage medium storing operation improvement assistance program, operation improvement assistance device, and operation improvement assistance method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |